We strive to build an interdisciplinary team working at the intersection of neuroscience, health science, and computer science. Our core values are:
- Innovation: consistently question the novelty, creativity, and practicality of our solution; our solution should be built upon a rigorous review of past literature and existing solutions.
- Industry-relevant: Our solution should be industry-relevant and have clear demands, mostly coming in the form of applications or systems and when possible, having identified industrial partners.
- Experimental: We value scientific rigor, focusing on researching under strong scientific grounds and conducting sound experiments that provide definitive and repeatable findings.
- Computational: Our scientific nature is to use algorithms, mathematical models, strong theoretical background, and strong coding skills.
People are usually not conscious nor have the tools to know about their general health and well-being. Our lab focuses on understanding the physiological (e.g., EEG, ECG, NIR, Raman) or physical profile (e.g., ultrasound, faces) of human health using advanced signal/image processing (e.g., blind source separation, Riemannian geometry, wavelet neural networks) and deep learning (e.g., transfer learning, unsupervised learning, GANs, attention mechanisms, graphs, meta-learning).
To achieve that, our lab contributes to the Human Sensing vision which is overarching software systems (hardware and software) that contain many key modules to be implemented.
This project envisions two categories of modules: (1) cross-modules, focusing on modules that work across different tasks, and (2) application modules, focusing on delivering specific features.
- Acquisition - enables users to simultaneously connect different sources of physiological data, e.g., EEG, ECG, EMG, spectroscopy.
- Artifact removal - enables auto-preprocessing of incoming signals, e.g., notch filter, bandpass filters, EOG/ECG artifact removals.
- Model - enables the auto search of optimal models for regression and classification tasks
- Visualization -enables interactive data reporting
- Stress submodules - stress is one of the most important predecessors of both mental and physical diseases. We seek to develop a cost-effective efficient stress wearable + mobile/cloud platform that can capture stress in real-time and send this data to the platform for recommendations and behavioral interventions. Works with the Relaxation module.
- Glucose submodules - I am a huge fan of the keto diet and intermittent fasting. Abundant evidence shows that sugar is the number one cause of cancers, as well as other critical diseases. The only way we can measure glucose is to prick our fingers. We seek to develop a portable/wearable glucose measuring tool (e.g., Raman spectroscopy) that can measure glucose non-invasively.
- Relaxation module - this module works with the Stress submodules as an intervention (e.g., meditations or neurofeedback training) to gradually train users to learn to lower their stress levels over time.
- Emotion - with incoming signals from various sources in real-time, it outputs the emotional and cognitive profile of users, e.g., valence, arousals.
- Personality module - this module aims to provide a quick, objective personality assessment using physiological signals, as compared to questionnaires.
- Input and Output module - this module aims to provide some interactive tasks (e.g., brain spellers, gaming), by leveraging incoming physiological signals as input as well as for adaptive purposes.